fix recurrent sac

ml-agents/mlagents/trainers/tests/tensorflow/test_simple_rl.py

 
 @pytest.mark.parametrize("use_discrete", [True, False])
 def test_recurrent_sac(use_discrete):
-    step_size = 0.2 if use_discrete else 0.5
+    step_size = 0.2 if use_discrete else 0.8
     env = MemoryEnvironment(
         [BRAIN_NAME], use_discrete=use_discrete, step_size=step_size
     )
         SAC_TF_CONFIG,
         hyperparameters=new_hyperparams,
         network_settings=new_networksettings,
-        max_steps=5000,
+        max_steps=4000,
         framework=FrameworkType.TENSORFLOW,
     )
     _check_environment_trains(env, {BRAIN_NAME: config})